Arduino’s Long-Awaited Improved WiFi Shield

Announced at the 2014 Maker Faire in New York, the latest Arduino WiFi shield is finally available. This shield replaces the old Arduino WiFi shield, while providing a few neat features that will come in very handy for the yet-to-be-developed Internet of Things.

While the WiFi Shield 101 was announced a year ago, the feature set was interesting. The new WiFi shield supports 802.11n, and thanks to a few of Atmel’s crypto chip offerings, this shield is the first official Arduino offering to support SSL.

The new Arduino WiFi Shield 101 features an Atmel ATWINC1500 module for 802.11 b/g/n WiFi connectivity. This module, like a dozen or so other WiFi modules, handles the heavy lifting of the WiFi protocol, including TCP and UDP protocols, leaving the rest of the Arduino free to do the actual work. While the addition of 802.11n  will be increasingly appreciated as these networks become more commonplace, the speed offered by ~n isn’t really applicable; you’re not going to be pushing bits out of an Arduino at 300 Mbps.

Also included on the WiFi shield is an ATECC508A CryptoAuthentication chip. This is perhaps the most interesting improvement over the old Arduino WiFi shield, and allows for greater security for the upcoming Internet of Things. WiFi modules already in the space have their own support for SSL, including TI’s CC3200 series of modules, Particle‘s Internet of Things modules, and some support for the ESP8266.

45 thoughts on “Arduino’s Long-Awaited Improved WiFi Shield

  1. Cool and everything, but, I get the feeling the arduino boards are becoming less relevant for two reasons:

    1. Form factor – this original size form factor is fairly hefty for a lot of applications
    2. Cost – by comparison to other options out there the arduino shields are just too expensive

    Since I found out you can use the arduino ide with the esp8266 I can’t imagine going back!

    1. There’s another reason:
      They’re in bed with Atmel, so of course they’re using an Atmel chip here.
      This greatly reduces their usable portfolio of chips.
      The Arduino Due is a good example. Instead of using an ARM M3 or M4 with 5V IO voltage (yes, they exist), or at least 5V tolerant inputs, they had to use Atmel chips, that only run on 3.3V.
      This generated a lot of compatibility problems, as there were (and still are) a whole universe of 5V based shields.

      One example that springs into mind are the reprap 3D printer RAMPS boards. They’re originally designed for the 5V Arduino Mega, but the Mega is nowadays too limited, so people are moving to 32bit controllers. A RAMPS version for the Due exists, but never became that much of a standard as the old ones, because as it was necessary to move to 3.3V anyways, the designers also looked for other micros besides the Atmel variants.
      And that’s why Arduino is losing it’s relevance in this segment.

      1. Have you got any suggestions for some fully 5v tolerant (adc aswell as gpio) arm chips?

        I’ve got the TI Hercules board based on TMS570 and I’m looking at the atsamc21 but I cant seen to see any lower cost chips.

      1. I’ve just started using an Arduino but is developing things that will use Wi-Fi hard? I’m not planning on doing anything wi-Fi based anytime soon, but would like to know anyway. Wi-Fi seems like a whole different universe. Thanks.

        1. I have had the same thought (Arduino form factor with an ESP8266 on it) many times and am so glad someone has done it.

          It makes a great dev/test paltform for the ESP and can be a drop in replacement for many existing Arduino projects with many more times Memory and CPU grunt available with Wifi to boot.

          My first Arduino WiFi shield cost me $90 and I really never used it, such a waste of money!

          1. It’s absurd that a shield can cost more than a ful openwrt capable wifi module with 2 ethernet ports on it, which being openwrt will do whatever TLS is needed just by updating it. The module for these are under $10 US, a breakout with the ethernet jacks and a serial port is a little over $20 – http://wiki.openwrt.org/toh/hilink/hlk-rm04

            Wont have the support of the yun out of the box, but IMO that thing is way overpriced.

  2. The layout of that board looks like the parts were dragged manually into a convenient location on the PCB and the auto router was left to handle the traces.

    Sure it can work just fine but it looks terrible. No thought into appropriate trace width, no thought for proper componenet placement or anything..

    PCBs can be a work of art in themselves and it doesn’t take a lot of work to have a board that looks great!

      1. Who said I was Anti-Arduino?

        I have a pile of them.. Uno’s, Leonardo’s, Pro Minis.. They are neat little boards for what they do..

        All I am saying is, perhaps spend more than 30 seconds laying out a board and auto-routing it… Have some pride in your PCBs!

      2. No they’re not, that routing is pretty bad, first thing I noticed actually. All the traces are the same width and clumped in odd ways, also no fill on the top layer? Probably not necessary since the wifi is actually on a daughterboard.

        They took an off-the-shelf IoT radio and put it on a badly routed arduino breakout board.

    1. This is a pet peeve of mine. The early Arduino boards were pitifully amateurish efforts. Poor routing casts doubt on the rest of the design, and there were a few electrical blunders as well.

      But that said, can you point to any specific aspect of the WiFi Shield that’s definitively bad RF design practice? The board is evidently four layers, with buried power and ground planes. Trace width looks appropriate to me: 5V power is sized wide, signals are narrow (free filtering). Traces generally look as short and direct as possible. The shield looks like any commercial product where the designer didn’t put effort into making curvy traces.

      We could do a more detailed analysis if the source files were available. What’s up with that, Arduino.cc?

  3. That only has single chain and 2.4GHz so its not going to be getting 300megabit anyway, just 72 on a 20 MHz wide channel. Still nice to not have to allow legacy devices on the network slowing it down like the stupid CC3000 based stuff.

      1. I wouldnt do anything with it that would generate more than that, it was in reference to the 300 in the post. But not having G or B devices on your 2.4GHz lan makes a huge difference to thruput since it doesnt have to do all the G protection stuff. Is the difference between 1080 netflix working and not working on a chromecast when I tried it last.

        1. The slowest device dictates the speed of the whole WiFi. As simple as that. This is the reason why one doesn’t want any clients being so far away from the AP that they connect with 10Mbps, too.

          1. No it is not as simple as that…that’s just plain untrue. Broadcasts need to be sent at the rate of the slowest device, but point to point comms will still occur at the speed negotiated between each individual device and the AP.

  4. I got one of those twinkie chips as a free sample from Atmel a while back. I intended on fiddling with it and seeing if I could get it working with some-duino-or-another. I never got around to it though. Maybe it’ll be a little easier now, since they’ve done the hard work. Now to get 40-some projects done and out of the way… Well, at least the one important one right now ;)

    1. not everyone wants a truly directly connected iot device.

      for me, I always want them behind a nat. that gives ‘free’ protection against inbound conect attempts, which, imho, should never happen in an iot network.

    2. NAT would only block traffic going in from the internet. A wireless device is not protected from direct hacking without a Faraday cage. An attacker can use a directional antenna, LNA etc. to access the device from a distance.

    1. let me ask you: if you are going to design a product to ship and depend on, are you going to pick a chinese chip that is mostly a hack or are you going to pick a module from a company like atmel that will be around, will support their module and is probably a better chip designer, overall, anyway?

      I still don’t know much about the esp company and if they are going to be around for the long run; if they will embed crap in the chip or backdoors, if they will change the chip and not tell anyone and still call it the same chip, etc.

      in short, I have esp modules to hack on for fun, but I would NEVER use this in a shipping product. no friggin way. I need to be sure of the company and that the product will be around, the same chip unchanged and be supported for years. I just cannot feel confident about esp’s from china.

  5. The whole idea that Arduino caught on was for fast prototyping. Just add what they want with a stackable board. A lot of times keeping the whole board in a final product. Which gets expensive quick if you add things like WiFi.

    For some of us though we are less put off by doing more of the hardware connections. Especially if you see you’re going to have to eventually anyway. We see prototyping as eventually getting down to minimal components. So we just start small and cheap to begin with. Our time and cost is in developing a circuit board in the end.

    I have to agree it’s too pricey for just adding WiFi. Just as well off using the HC-05, 06, or even the ESP8266 if you just want to add WiFi. That is if you find it necessary to use the larger Arduino boards. But I see that they want to stick with the plug-n-play theme. That’s a lot of money wrapped up for WiFi.

    What the ESP8266 lacks is ADC input which only the ESP-12 has 1. So if you need more than one you’re going to come up short. If you need any the others don’t have any. If you need more than 1 then you can still use it for your Wifi for your Arduino. But then that opens the argument for wasting a 32-bit processor. But not really if it’s that cheap to get WiFi and you have to have the ADC.

    A lot of my use cases are for Ham radio or handheld instruments do not need WiFi . But if I wanted it’s very cheap to add it. Who knows I might find a reason that they could benefit from WiFi connection. I could add a lot for the price of that one WiFi board.

    1. It’s not terrible hard to design an in analog MUX chip, or even just add a breakout board if you aren’t designing into your own circuit board (https://www.sparkfun.com/products/9056).

      I get the point of the shields for people that are looking for plug and play (there are a lot of artists that don’t really know electronics but what to have a uC controlled art piece).

      I am slowly learning about the simple hardware people often avoid like MUX/DEMUX chips or say using shift registers to drive led displays.

      I think it is interesting to see the choices made by the Arduino team. Not sure about no IPv6, but in the end, it’ll probably go on a private IPv4 network anyhow.

      1. Its even easier to add an 8 pin spi or I2C external A/D, you can have 16 or even 24 bits for really cheap. That’s much better than the 12bits (10 bits for arduino) found in most microcontrollers

  6. I can’t quite tell; is there a ufl connector there? I see a pcb antenna but if this is inside a box, that onboard antenna will be blocked. if this is meant for iot then it HAS to allow an external antenna, even if just a stubby one.

  7. Can someone please educate me on WiFi options — I can see using Particle Photons for future network-connected projects, but in the meantime, I have a bunch of Arduino Uno and Nano clones that I would like to get WiFi-connected. I understand the ESP8266 is out there but what is the most turn-key way of using them to connect to my router? I don’t want to spend a lot of time fiddling with the WiFi connectivity.

  8. So the ATECC508A CryptoAuthentication chip seems to be using EC crypto only (256bit, roughly equivalent to 3072 bit RSA key for traditional computing), which is already starting to become obsolete because of possibility of quantum computing attacks – 224bit EC key is about half as strong as 2048-bit RSA key for quantum computer.. in terms of qubits, at least.
    Of course quantum computers are not widely available yet, but this is just the theoretical research so far — it’s quite possible new attacks would surface, so all your IoT devices may become really vulnerable (and since devices get deployed and forgotten about for many years at a time…)

    Just a thought, not an argument really…

  9. Seems the Arduino guys are too busy with the brand fight not to note the quick progress done by the ESP8266 community on arduino IDE, I’ve one of those cheap 6$ boards with 10 GPIO ADC etc, and I’m amazed how quick and cheap is to develop with them, If you have an good router and engage your IOT devices to somenthing like Domoticz runnin on a Raspi or a NAS you dont have to take care on encryption sonce all the traffic is in-doors, however I’ll re-consider this board for some ROAMING device conecting to public wifi or outside the network firewall to the bare internet.

Leave a Reply to Dewgenenny Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.